In 2020: (1) Platforms were increasingly specific in their attributions – more often naming a govt, firm, or actor behind an operation. Based on our coding, approximately 76% of takedowns were attributed to a specific actor in 2020, compared to 62% in 2019 and 47% in 2018.
There are costs to the platforms for making specific attributions. For example, when Twitter suspended a network linked to the youth branch of Turkey’s ruling party, the office of the presidency responded...
(2) Disinformants innovated. We saw more Twitter networks leverage handle switching - using the tactic to create accounts with >10k followers at times - with no easy way to figure out their identity in a previous life.
We also saw the spread of AI-created pro pics, a way to evade reverse image searches. ≥7 ops in 2020 (including a Cuba-linked network) used AI pics. @Graphika/@DFRLab analyzed the first takedown that deployed the tactic “at scale” in 2019. graphika.com/reports/operat…
(3) Political actors are increasingly outsourcing disinfo. In 2020, FB/Twitter attributed ≥15 operations to private firms (eg PR firms). Outsourcing has benefits: expertise, plausible deniability. But there also may be principal-agent problems at play. lawfareblog.com/outsourcing-di…
(4) All countries were not targeted equally. When it comes to foreign operations in 2020, the U.S. was targeted in 14 takedowns, the UK in 4, and Egypt in 5. Note that this doesn’t reflect the universe of info ops, since our data is limited to platform takedowns.
(5) Fake news outlets remain a popular tactic, but as platform detection of covert social media networks improves, actors have further incentive to invest in off-platform news websites. We’ve seen the creation of “original” outlets and typosquatting to impersonate real entities.
Big thanks to @EliasGroll for superb editing. It’s especially exciting to publish w/@BrookingsInst as a former intern. The summer I spent working in @BrookingsFP for @ChhabraT propelled my interest in security, technology, and international order.
We hope the piece will contribute to ongoing efforts to learn broader lessons from takedowns and researcher reports, and inform policymakers’ thinking about how to combat malign influence in 2021.
• • •
Missing some Tweet in this thread? You can try to
force a refresh